humanizer

What problem does it solve? Academic drafts written or assisted by AI often contain recognizable patterns—significance inflation, hedge-stacking, formulaic openers, and boilerplate transitions—that make the prose read as machine-generated and weaken its credibility with reviewers. This Skill audits a draft for those tells and rewrites it into direct, scholarly prose. ## Core Features & Use Cases - AI Tell Audit: Scans a draft and reports a bulleted list of detected AI patterns, from em dash overuse to passive evasion and rule-of-three structures. - Academic-Specific Rewrites: Targets paper-level issues like contribution inflation, formulaic abstracts, robustness boilerplate, literature padding, and roadmap filler. - Domain-Aware Revision: Includes TDA/social-science-specific guidance, such as naming topological features and persistence values instead of vague claims. - Use Case: Before submitting a paper draft, run the pass on the full manuscript to receive a diagnosis of AI-sounding passages plus a revised version ready for review. ## Quick Start Run the humanizer pass on my attached paper draft and return the audit of AI writing patterns followed by the revised text.